Presentation is loading. Please wait.

Presentation is loading. Please wait.

Always race to sleep? i.e. how we managed to confuse ourselves by talking about two kinds of race to sleep.

Similar presentations


Presentation on theme: "Always race to sleep? i.e. how we managed to confuse ourselves by talking about two kinds of race to sleep."— Presentation transcript:

1 Always race to sleep? i.e. how we managed to confuse ourselves by talking about two kinds of race to sleep

2 Always race to sleep? i.e. how we managed to confuse ourselves by talking about two kinds of race to sleep Single node – amortize static power cost Many nodes – minimize parallelization overhead

3 1. Background: Common system power and work rate behaviors 2. Analysis: Required conditions for race to sleep 3. Empirical data: When are those conditions met in Hadoop?

4 Two common behaviors for system power Resource proportional Not resource proportional

5 Three common behaviors for system work rate Linear speed up Parallelization overhead Bottleneck elsewhere

6 1. Background: Common system power and work rate behaviors 2. Analysis: Required conditions for race to sleep 3. Empirical data: When are those conditions met in Hadoop?

7 Work rate

8 Power

9 Work rate Power Power efficiency = work rate / power

10 Work rate Power Power efficiency = work rate / power

11 Work rate Power Power efficiency = work rate / power Yes Time benefit, no energy benefit Yes Increasing efficiency Race to sleep? i.e. operate at highest work rate? No Decreasing efficiency Somewhat Turning point exists Energy benefit, no time benefit Yes and no

12 Work rate Power Power efficiency = work rate / power Yes Increasing efficiency Race to sleep? i.e. operate at highest work rate? No Decreasing efficiency Somewhat Turning point exists Yes and no Time benefit, no energy benefit Energy benefit, no time benefit

13 Work rate Power Power efficiency = work rate / power Race to sleep? i.e. operate at highest work rate? Go faster if % increase in work rate ≥ % increase in power Go slower otherwise

14 Work rate Power Power efficiency = work rate / power Race to sleep? i.e. operate at highest work rate? Required condition for race to sleep Go faster if % increase in work rate ≥ % increase in power Go slower otherwise

15 Work rate Power Power efficiency = work rate / power Race to sleep? i.e. operate at highest work rate? e.g. Old work rate = A New work rate = 1.1A Old power = B New power = 1.05A Old power efficiency = A / B New power efficiency = (1.1 / 1.05) × (A / B) = (1.1 / 1.05) × old power eff. Required condition for race to sleep Go faster if % increase in work rate ≥ % increase in power Go slower otherwise

16 1. Background: Common system power and work rate behaviors 2. Analysis: Required conditions for race to sleep 3. Empirical data: When are those conditions met in Hadoop?

17 Hadoop sort 10GB terasort format HDFS read 10GB HDFS write 10GB Hadoop shuffle 10GB

18 Work rate Hadoop sort 10GB terasort format HDFS read 10GB HDFS write 10GB Hadoop shuffle 10GB

19 Work rate Hadoop sort 10GB terasort format HDFS read 10GB HDFS write 10GB Hadoop shuffle 10GB Power efficiency

20 Work rate Hadoop sort 10GB terasort format HDFS read 10GB HDFS write 10GB Hadoop shuffle 10GB Power efficiency Race to sleep No Yes No Yes

21 That was multi-node power efficiency Single-node power efficiency is a different picture

22 Always race to sleep?

23 Maybe the question should be Always use as much resources as possible?

24 Always race to sleep? Maybe the question should be Always use as much resources as possible? Take away: Single node – amortize static power cost (awake nodes should race to sleep) Many nodes – minimize parallelization overhead (as few nodes awake as possible) Increase resource if resulting % work rate increase ≥ % power increase

25 Work rate Power Power efficiency = work rate / power Yes Increasing efficiency Race to sleep? i.e. operate at highest work rate? No Decreasing efficiency Somewhat Turning point exists Yes and no Time benefit, no energy benefit Energy benefit, no time benefit

26 Other junk …

27 Power efficiency = energy efficiency

28 =

29

30


Download ppt "Always race to sleep? i.e. how we managed to confuse ourselves by talking about two kinds of race to sleep."

Similar presentations


Ads by Google